INDEX

Explanations

mentions of health conditions, particularly malignancies

oai_token-act-pair · gpt-3.5-turbo

terms related to malfunctions or negative conditions

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

BOOK

-0.82

æĸ¹

-0.74

 FACE

-0.72

ITED

-0.72

DragonMagazine

-0.71

 Hobby

-0.71

zzo

-0.70

 Carbuncle

-0.69

hetti

-0.69

 Solitaire

-0.68

POSITIVE LOGITS

mal

1.15

 vulner

1.00

adies

0.96

colm

0.91

ignant

0.88

ciating

0.79

mal

0.79

 challeng

0.79

practice

0.78

querade

0.78

Activations Density 0.008%

mentions of health conditions, particularly malignancies

terms related to malfunctions or negative conditions

No Comments

No Known Activations